site_archiver@lists.apple.com Delivered-To: darwin-kernel@lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:to :subject:in-reply-to:mime-version:content-type :content-transfer-encoding:content-disposition:references; bh=Z3SzDFiOOFewDgl8hn9ewcEn56yW7g/HmAg97x1rL94=; b=DcSGq+O1fY297eyYgG3dN3PG/9wx7ihmZHlHPuyQJ8XGEmrXZpUo2KDZ/r/lSqb3ps uI7EwVYf5Rj/vLdsh3cyoprjKAV/SSLLYrTl/bVcTelTG2mClrdCBpktSqUvzIU2ZxbK UYFGa6Aca+KYHHXG5rC+XnXPvFok9g9bZmlNQ= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:in-reply-to:mime-version :content-type:content-transfer-encoding:content-disposition :references; b=cw4u71RSzuWr1Y9gkT2eMW15QLS90TyYdsjksL0N1oiP8Dn3WtdoHdM9hK+hhWhYO0 Uab3ZbHKdFJoI88QhLMgj1puq9slk86ZpD1E97oHekqq+qos30KTrCQWLvXh/gar6Afy 7pVWHrl5Ys0XF4DTFQ71u+gEJZDcw3x0hqANY= doh, I was rushing and left two unused definitions in the snippet. The definitions of kfs_event_arg_t, kfs_event are obviously unnecessary. sorry for the inconvenience. On Thu, Aug 7, 2008 at 1:30 PM, Jernej Azarija <azi.stdout@gmail.com> wrote:
Hello,
I've developed a small application using the fsevents mechanism. The application froze the machine within a few moments of usage.
I've went on to inspect the problem and trim the whole source to ~70 lines of code reproducing the panic. I apologize for pasting the code here :
== #include <sys/types.h> #include <sys/ioctl.h> #include <unistd.h> #include <fcntl.h> #include <stdio.h>
#include "fsevents.h"
static int fsevent_fd;
typedef struct kfs_event_arg { u_int16_t type; u_int16_t len; union { struct vnode *vp; char *str; void *ptr; int32_t int32; dev_t dev; ino_t ino; int32_t mode; uid_t uid; gid_t gid; } data; } kfs_event_arg_t;
typedef struct kfs_event { int32_t type; pid_t pid; kfs_event_arg_t args[FSE_MAX_ARGS]; } kfs_event;
static int8_t events[FSE_MAX_EVENTS];
int main() {
int fd,ret; struct fsevent_clone_args fca; int fsevent_fd;
int i; for (i = 0; i < FSE_MAX_EVENTS; i++) /* BUG */ events[i] = FSE_REPORT; /* BUG */
fd = open("/dev/fsevents", O_RDONLY);
if (fd == -1) { return 1; } fca.event_list = (int8_t *)events; fca.num_events = sizeof(events)/sizeof(int8_t); fca.fd = &fsevent_fd;
ret = ioctl(fd, FSEVENTS_CLONE, (char *)&fca);
if (ret == -1) close(fd); return 1; close(fd);
puts("Sleeping a little"); sleep(60*5);
return 0; } == For the code to compile, the "fsevents.h" header must be present and is therefore included as an attachment.
The interesting part is, that removing the lines marked with /*BUG*/ (ie, the for loop) fixes the problem! I'm clueless about what/why is causing the panic crash. I'm even more confused after looking at panic.log reporting division by zero as the cause of error.
=== Thu Aug 7 13:12:57 2008 panic(cpu 1 caller 0x001A49CB): Unresolved kernel trap (CPU 1, Type 0=divide error), registers: CR0: 0x80010033, CR2: 0xa001001c, CR3: 0x00dc9000, CR4: 0x000006e0 EAX: 0x00000001, EBX: 0x03aa57c4, ECX: 0x00000000, EDX: 0x00000000 CR2: 0x2536bb1c, EBP: 0x2536bc58, ESI: 0x00469d64, EDI: 0x00469d60 EFL: 0x00010246, EIP: 0x001ef7ce, CS: 0x00000004, DS: 0x0172000c
Backtrace, Format - Frame : Return Address (4 potential args on stack) 0x2536b988 : 0x128d0d (0x3cc65c 0x2536b9ac 0x131f95 0x0) 0x2536b9c8 : 0x1a49cb (0x3d2a94 0x1 0x0 0x3d21e8) 0x2536bad8 : 0x19b3a4 (0x2536bae8 0x2536bf48 0xe 0x48) 0x2536bc58 : 0x1e8627 (0x8 0x2536bf48 0x1 0x3c67528) 0x2536bca8 : 0x1da43b (0x3c67528 0x2536be30 0x2536bf48 0x2536bf48) 0x2536bf28 : 0x1da486 (0x1 0x423f7f0 0x2536bf58 0x199489) 0x2536bf58 : 0x37b300 (0x4293bb8 0x41b7a98 0x41b7adc 0x0) 0x2536bfc8 : 0x19b77e (0x423f7ec 0x0 0x19e0b5 0x423f9e0) No mapping exists for frame pointer Backtrace terminated-invalid frame pointer 0xbfffb5c8
Kernel version: Darwin Kernel Version 8.11.1: Wed Oct 10 18:23:28 PDT 2007; root:xnu-792.25.20~1/RELEASE_I386 ===
Anyone clued about what is going on? If there is some other information I could provide you with, please tell.
Thanks,
Jernej
-- http://nd47.coderock.org/~azi _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.a... This email sent to site_archiver@lists.apple.com